pith. sign in

arxiv: 2605.18769 · v1 · pith:N2M6KDFBnew · submitted 2026-04-14 · 💻 cs.IR · cs.AI· cs.CL

ClusterRAG: Cluster-Based Collaborative Filtering for Personalized Retrieval-Augmented Generation

Gibson Nkhata , Uttamasha Anjally Oyshi , Quan Mai , Susan Gauch This is my paper

Pith reviewed 2026-05-21 00:58 UTC · model grok-4.3

classification 💻 cs.IR cs.AIcs.CL
keywords personalized RAGcollaborative filteringdensity-based clusteringuser profilesretrieval-augmented generationLaMP benchmarkcluster-based retrieval
0
0 comments X

The pith

Clustering users by profile documents and retrieving from similar groups improves personalized RAG performance.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper proposes ClusterRAG to make personalized retrieval-augmented generation more accurate by adding collaborative signals from users with similar interests. It represents each user by their profile documents, applies density-based clustering to form coherent groups, and retrieves information first at the cluster level then with fine-grained document ranking. Experiments on the LaMP benchmark show that combining the target user's own profile with profiles from the most similar users produces the strongest results across tasks. The method works alongside various dense retrievers and both fine-tuned and zero-shot language models.

Core claim

ClusterRAG represents users through their profile documents, organizes users into semantically coherent clusters using density-based clustering, and performs retrieval at both the cluster and document levels via cluster-level similarity and fine-grained ranking. Extensive experiments on the LaMP benchmark demonstrate that jointly leveraging the target user's profile and profiles from top similar users consistently yields the best performance across diverse tasks.

What carries the argument

Density-based clustering applied to user profile documents, enabling combined cluster-level similarity search and document-level ranking for collaborative retrieval.

If this is right

  • Jointly using the target profile and similar-user profiles delivers the highest scores on LaMP tasks.
  • The system integrates directly with multiple dense retrievers and rankers without modification.
  • Performance gains hold when the underlying language model is either fine-tuned or zero-shot.
  • Cluster organization reduces the need to scan every user profile during retrieval.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same clustering step could improve group-aware recommendation in non-RAG settings such as content suggestion.
  • Scaling tests on larger user populations would show whether cluster coherence persists outside the LaMP collection.
  • Semantic user clusters may supply a general mechanism for reducing retrieval cost in any profile-driven generative system.

Load-bearing premise

Density-based clustering on profile documents produces semantically coherent user groups whose collaborative signals reliably improve retrieval for the target user.

What would settle it

An ablation on the LaMP benchmark that disables the cluster-based component and finds no gain or a drop in performance compared to using only the target user's profile.

Figures

Figures reproduced from arXiv: 2605.18769 by Gibson Nkhata, Quan Mai, Susan Gauch, Uttamasha Anjally Oyshi.

Figure 1
Figure 1. Figure 1: Overview of the ClusterRAG framework. ciency and generalization. These properties make ColBERTv2 well-suited for robust similarity esti￾mation between user profiles. Cluster-Level Similarity Ranking. This step aims at restricting similarity computation to clus￾ter members to improve robustness and scalability by focusing comparisons on behaviorally consis￾tent cohorts. For each cluster C, we construct an i… view at source ↗
Figure 2
Figure 2. Figure 2: Retrivers’ ROGUE-1 scores on LaMP-5. ment frequency (TF-IDF), BM25 (Robertson et al., 1995), (4) a heuristic retriever that ranks documents solely based on temporal proximity to the query time, favoring the most recently published docu￾ments, Recency, and (5) a non-informative base￾line that samples documents uniformly at random, Random. We provide retriever-agnostic design re￾sults in [PITH_FULL_IMAGE:fi… view at source ↗
read the original abstract

Personalized Retrieval-Augmented Generation (RAG) relies on accurately selecting user-relevant documents. In practice, existing RAG approaches often suffer from high retrieval costs and overlook that collaborative signals from similar users can enhance personalized generation for the current user. We propose ClusterRAG, a Cluster-Based Collaborative Filtering for Personalized Retrieval-Augmented Generation. ClusterRAG represents users through their profile documents, organizes users into semantically coherent clusters using density-based clustering, and performs retrieval at both the cluster and document levels via cluster-level similarity and fine-grained ranking. Extensive experiments on the LaMP benchmark demonstrate that jointly leveraging the target user's profile and profiles from top similar users consistently yields the best performance across diverse tasks. Further analysis shows that ClusterRAG integrates seamlessly with different dense retrievers and rankers, and remains effective when paired with both fine-tuned and zero-shot language models.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The paper proposes ClusterRAG, which represents users by their profile documents, applies density-based clustering to form semantically coherent groups, and performs retrieval using both cluster-level similarity and fine-grained document ranking. It incorporates collaborative signals from similar users' profiles together with the target user's profile. Experiments on the LaMP benchmark are reported to show that this joint approach yields the best performance across diverse tasks, with compatibility to multiple dense retrievers and both fine-tuned and zero-shot language models.

Significance. If the clustering step produces meaningful groups and the observed gains are attributable to collaborative signals rather than simply larger retrieval pools, the method could offer a practical enhancement to personalized RAG by structuring user similarity. The reported compatibility with varied retrievers and LLMs is a positive practical feature.

major comments (2)
  1. [Abstract] Abstract (method paragraph): the central claim that density-based clustering on profile documents produces semantically coherent user groups whose collaborative signals improve retrieval is load-bearing, yet no cluster-quality metrics (e.g., silhouette score, intra-cluster coherence) or ablation against non-clustered kNN profile retrieval are described; without these, gains could arise from simply retrieving over more profiles rather than from the clusters themselves.
  2. [Results] Results section (performance tables): the statement of 'consistent best performance' across tasks requires explicit reporting of statistical tests, error bars, and direct comparison to a non-clustered collaborative baseline (e.g., top-k similar users selected by embedding similarity without density clustering) to isolate the contribution of the clustering step.
minor comments (2)
  1. Specify the exact density-based algorithm (DBSCAN, HDBSCAN, etc.) and the procedure for selecting or tuning its parameters (eps, min_samples, distance metric).
  2. Ensure figure captions and table footnotes clearly indicate which retriever and LLM variant are used in each row so that the integration claims can be verified at a glance.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback on ClusterRAG. The comments highlight important ways to strengthen the evidence for the clustering component and the statistical robustness of our results. We address each point below and will incorporate revisions to improve the manuscript.

read point-by-point responses

  1. Referee: [Abstract] Abstract (method paragraph): the central claim that density-based clustering on profile documents produces semantically coherent user groups whose collaborative signals improve retrieval is load-bearing, yet no cluster-quality metrics (e.g., silhouette score, intra-cluster coherence) or ablation against non-clustered kNN profile retrieval are described; without these, gains could arise from simply retrieving over more profiles rather than from the clusters themselves.

    Authors: We agree that cluster-quality metrics and a direct ablation against non-clustered kNN retrieval would better isolate the contribution of density-based clustering. While our primary evaluation focuses on end-task performance on LaMP, we will add silhouette scores and intra-cluster coherence metrics in the revised manuscript. We will also include an ablation comparing ClusterRAG to a kNN baseline that selects top-k similar users by embedding similarity without density-based clustering. This will clarify that performance gains arise from the semantically coherent groups rather than simply from a larger retrieval pool. revision: yes

  2. Referee: [Results] Results section (performance tables): the statement of 'consistent best performance' across tasks requires explicit reporting of statistical tests, error bars, and direct comparison to a non-clustered collaborative baseline (e.g., top-k similar users selected by embedding similarity without density clustering) to isolate the contribution of the clustering step.

    Authors: We acknowledge that explicit statistical tests, error bars, and the non-clustered baseline are needed for rigor. In the revised version, we will report standard deviations across multiple runs as error bars and include paired statistical significance tests (e.g., t-tests) for the observed improvements. As noted in our response to the abstract comment, we will also add the direct kNN-without-clustering baseline to the performance tables to demonstrate the specific benefit of the density-based clustering step. revision: yes

Circularity Check

0 steps flagged

No circularity: empirical method on external benchmark

full rationale

The paper describes an empirical clustering-based collaborative filtering approach for personalized RAG and evaluates it directly on the external LaMP benchmark. No equations, derivations, or first-principles claims are present that reduce performance metrics to parameters fitted from the same data or to self-referential definitions. The central results are experimental comparisons, not closed-form predictions that collapse to the inputs by construction. Self-citations, if any, are not load-bearing for the reported gains.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

Review is based only on the abstract; the central claim rests on the domain assumption that user-profile clustering captures useful collaborative signals and on the existence of tunable clustering parameters whose values are not specified here.

free parameters (1)
  • density-based clustering parameters
    DBSCAN-style clustering requires at least one distance threshold and minimum cluster size; these must be chosen or tuned and directly affect which users are grouped together.
axioms (1)
  • domain assumption Users with similar profile documents share relevant collaborative signals that improve document retrieval for the target user.
    This premise is invoked when the method moves from individual profiles to cluster-level retrieval.

pith-pipeline@v0.9.0 · 5686 in / 1314 out tokens · 82424 ms · 2026-05-21T00:58:05.081967+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

91 extracted references · 91 canonical work pages · 5 internal anchors

  1. [1]

    Advances in Neural Information Processing Systems 33 (NeurIPS 2020) , year=

    Retrieval-Augmented Generation for Knowledge-Intensive NLP , author=. Advances in Neural Information Processing Systems 33 (NeurIPS 2020) , year=

    work page 2020

  2. [2]

    2025 , eprint=

    A Survey of Personalization: From RAG to Agent , author=. 2025 , eprint=

    work page 2025

  3. [3]

    Proceedings of the 43rd International ACM SIGIR Conference on Research and Development in Information Retrieval , pages=

    ColBERT: Efficient and Effective Passage Search via Contextualized Late Interaction over BERT , author=. Proceedings of the 43rd International ACM SIGIR Conference on Research and Development in Information Retrieval , pages=. 2020 , doi=

    work page 2020

  4. [4]

    Colbertv2: Effective and efficient retrieval via lightweight late interaction

    Keshav Santhanam and Omar Khattab and Jon Saad-Falcon and Christopher Potts and Matei Zaharia. ColBERTv2: Effective and Efficient Retrieval via Lightweight Late Interaction. arXiv preprint arXiv:2112.01488. 2021

    work page arXiv 2021

  5. [7]

    ArXiv , year=

    A Survey on Retrieval-Augmented Text Generation for Large Language Models , author=. ArXiv , year=

    work page

  6. [8]

    Proceedings of the Seventeenth International Conference on Information, Process, and Knowledge Management , pages=

    ColBERT-Based User Profiles for Personalized Information Retrieval , author=. Proceedings of the Seventeenth International Conference on Information, Process, and Knowledge Management , pages=. 2025 , url=

    work page 2025

  7. [9]

    2024 , eprint=

    PersonaRAG: Enhancing Retrieval-Augmented Generation Systems with User-Centric Agents , author=. 2024 , eprint=

    work page 2024

  8. [10]

    Applied and Computational Engineering , volume =

    A Retrieval-augmented Generation Framework with Retriever and Generator Modules for Enhancing Factual Consistency , author =. Applied and Computational Engineering , volume =. 2025 , doi =

    work page 2025

  9. [14]

    International Journal of Data Science and Analytics , volume =

    An evaluation review of user similarity metrics in sparse collaborative filtering datasets , author=. International Journal of Data Science and Analytics , volume =. 2025 , doi =

    work page 2025

  10. [15]

    Proceedings of the Twenty-Sixth International Joint Conference on Artificial Intelligence,

    Hong-Jian Xue and Xinyu Dai and Jianbing Zhang and Shujian Huang and Jiajun Chen , title =. Proceedings of the Twenty-Sixth International Joint Conference on Artificial Intelligence,. 2017 , doi =

    work page 2017

  11. [16]

    Journal of Open Source Software (JOSS) , volume=

    hdbscan: Hierarchical density based clustering , author=. Journal of Open Source Software (JOSS) , volume=. 2017 , doi=

    work page 2017

  12. [18]

    Research on k-means Clustering Algorithm: An Improved k-means Clustering Algorithm , year=

    Na, Shi and Xumin, Liu and Yong, Guan , booktitle=. Research on k-means Clustering Algorithm: An Improved k-means Clustering Algorithm , year=

    work page

  13. [20]

    Transactions on Machine Learning Research , year=

    Unsupervised Dense Information Retrieval with Contrastive Learning , author=. Transactions on Machine Learning Research , year=

    work page

  14. [21]

    Okapi at TREC-3

    Robertson, Stephen and Walker, Steve and Jones, Susan and Hancock-Beaulieu, Micheline and Gatford, Mike , publisher =. Okapi at TREC-3. , url =. In Proceedings of the Third Text REtrieval Conference , year =

    work page

  15. [22]

    Akari Asai and Zeqiu Wu and Yizhong Wang and Avirup Sil and Hannaneh Hajishirzi , booktitle=. Self-. 2024 , address=

    work page 2024

  16. [23]

    2025 , eprint=

    HeteRAG: A Heterogeneous Retrieval-augmented Generation Framework with Decoupled Knowledge Representations , author=. 2025 , eprint=

    work page 2025

  17. [24]

    Data Intelligence , volume =

    Guo, Qian and Chen, Wei and Wan, Huaiyu , title =. Data Intelligence , volume =. 2021 , month =. doi:10.1162/dint_a_00104 , url =

    work page doi:10.1162/dint_a_00104 2021

  18. [25]

    ArXiv , year=

    RECOMP: Improving Retrieval-Augmented LMs with Compression and Selective Augmentation , author=. ArXiv , year=

    work page

  19. [28]

    Edward J Hu and yelong shen and Phillip Wallis and Zeyuan Allen-Zhu and Yuanzhi Li and Shean Wang and Lu Wang and Weizhu Chen , booktitle=. Lo. 2022 , url=

    work page 2022

  20. [29]

    Personal

    Thomas P Zollo and Andrew Wei Tung Siah and Naimeng Ye and Ang Li and Hongseok Namkoong , booktitle=. Personal. 2025 , url=

    work page 2025

  21. [31]

    2025 , eprint=

    Latent Inter-User Difference Modeling for LLM Personalization , author=. 2025 , eprint=

    work page 2025

  22. [32]

    2025 , eprint=

    Embedding-to-Prefix: Parameter-Efficient Personalization for Pre-Trained Large Language Models , author=. 2025 , eprint=

    work page 2025

  23. [33]

    ArXiv , year=

    A Survey of Controllable Learning: Methods and Applications in Information Retrieval , author=. ArXiv , year=

    work page

  24. [35]

    ArXiv , year=

    Think Before Recommend: Unleashing the Latent Reasoning Power for Sequential Recommendation , author=. ArXiv , year=

    work page

  25. [37]

    ArXiv , year=

    Test-Time Alignment for Tracking User Interest Shifts in Sequential Recommendation , author=. ArXiv , year=

    work page

  26. [38]

    ArXiv , year=

    QAGCF: Graph Collaborative Filtering for Q&A Recommendation , author=. ArXiv , year=

    work page

  27. [39]

    2025 , eprint=

    Collaborative Retrieval for Large Language Model-based Conversational Recommender Systems , author=. 2025 , eprint=

    work page 2025

  28. [40]

    ArXiv , year=

    Retrieval-Augmented Generation for Large Language Models: A Survey , author=. ArXiv , year=

    work page

  29. [42]

    2019 , eprint=

    BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding , author=. 2019 , eprint=

    work page 2019

  30. [43]

    2025 , eprint=

    A Systematic Survey of Prompt Engineering in Large Language Models: Techniques and Applications , author=. 2025 , eprint=

    work page 2025

  31. [44]

    2019 , eprint=

    Decoupled Weight Decay Regularization , author=. 2019 , eprint=

    work page 2019

  32. [45]

    2019 , eprint=

    PyTorch: An Imperative Style, High-Performance Deep Learning Library , author=. 2019 , eprint=

    work page 2019

  33. [46]

    2024 , eprint=

    Qwen2 Technical Report , author=. 2024 , eprint=

    work page 2024

  34. [48]

    PyTorch: an imperative style, high-performance deep learning library , year =

    Paszke, Adam and Gross, Sam and Massa, Francisco and Lerer, Adam and Bradbury, James and Chanan, Gregory and Killeen, Trevor and Lin, Zeming and Gimelshein, Natalia and Antiga, Luca and Desmaison, Alban and K\". PyTorch: an imperative style, high-performance deep learning library , year =. Proceedings of the 33rd International Conference on Neural Informa...

    work page

  35. [49]

    Advances in neural information processing systems , url =

    Pytorch: An imperative style, high-performance deep learning library , author=. Advances in neural information processing systems , url =

    work page

  36. [51]

    arXiv preprint arXiv:2406.00029 , year=

    Clustered Retrieved Augmented Generation (CRAG) , author=. arXiv preprint arXiv:2406.00029 , year=. 2406.00029 , archivePrefix=

    work page arXiv

  37. [52]

    Chi and Jeff Dean and Jacob Devlin and Adam Roberts and Denny Zhou and Quoc V

    Hyung Won Chung and Le Hou and Shayne Longpre and Barret Zoph and Yi Tay and William Fedus and Yunxuan Li and Xuezhi Wang and Mostafa Dehghani and Siddhartha Brahma and Albert Webson and Shixiang Shane Gu and Zhuyun Dai and Mirac Suzgun and Xinyun Chen and Aakanksha Chowdhery and Alex Castro-Ros and Marie Pellat and Kevin Robinson and Dasha Valter and Sha...

    work page

  38. [53]

    2025 , month = nov, url =

    Peter Steinberger , title =. 2025 , month = nov, url =

    work page 2025

  39. [57]

    ROUGE : A Package for Automatic Evaluation of Summaries

    Lin, Chin-Yew. ROUGE : A Package for Automatic Evaluation of Summaries. Text Summarization Branches Out. 2004

    work page 2004

  40. [62]

    Aleena Ahmad, Gibson Nkhata, Abdul Rafay Bajwa, Hannah Marsico, Bryan Le, and Susan Gauch. 2025. https://www.thinkmind.org/library/eKNOW/eKNOW_2025/eknow_2025_1_80_60025.html Colbert-based user profiles for personalized information retrieval . In Proceedings of the Seventeenth International Conference on Information, Process, and Knowledge Management, eKN...

    work page 2025

  41. [63]

    Akari Asai, Zeqiu Wu, Yizhong Wang, Avirup Sil, and Hannaneh Hajishirzi. 2024. https://openreview.net/forum?id=hSyW5go0v8 Self- RAG : Learning to retrieve, generate, and critique through self-reflection . In The Twelfth International Conference on Learning Representations, Vienna, Austria

    work page 2024

  42. [64]

    Hyung Won Chung, Le Hou, Shayne Longpre, Barret Zoph, Yi Tay, William Fedus, Yunxuan Li, Xuezhi Wang, Mostafa Dehghani, Siddhartha Brahma, Albert Webson, Shixiang Shane Gu, Zhuyun Dai, Mirac Suzgun, Xinyun Chen, Aakanksha Chowdhery, Alex Castro-Ros, Marie Pellat, Kevin Robinson, and 16 others. 2024. http://jmlr.org/papers/v25/23-0870.html Scaling instruct...

    work page 2024

  43. [65]

    Guanting Dong, Yutao Zhu, Chenghao Zhang, Zechen Wang, Ji-Rong Wen, and Zhicheng Dou. 2025. https://doi.org/10.1145/3696410.3714717 Understand what llm needs: Dual preference alignment for retrieval-augmented generation . In Proceedings of the ACM on Web Conference 2025, WWW '25, page 4206–4225, New York, NY, USA. Association for Computing Machinery

    work page doi:10.1145/3696410.3714717 2025

  44. [66]

    Wenqi Fan, Yujuan Ding, Liangbo Ning, Shijie Wang, Hengyun Li, Dawei Yin, Tat-Seng Chua, and Qing Li. 2024. https://doi.org/10.1145/3637528.3671470 A survey on rag meeting llms: Towards retrieval-augmented large language models . In Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, KDD '24, page 6491–6501, New York, NY,...

    work page doi:10.1145/3637528.3671470 2024

  45. [67]

    Markus Freitag and Yaser Al-Onaizan. 2017. https://doi.org/10.18653/v1/W17-3207 Beam search strategies for neural machine translation . In Proceedings of the First Workshop on Neural Machine Translation, pages 56--60, Vancouver. Association for Computational Linguistics

    work page doi:10.18653/v1/w17-3207 2017

  46. [68]

    Yunfan Gao, Yun Xiong, Xinyu Gao, Kangxiang Jia, Jinliu Pan, Yuxi Bi, Yi Dai, Jiawei Sun, Qianyu Guo, Meng Wang, and Haofen Wang. 2023. https://api.semanticscholar.org/CorpusID:266359151 Retrieval-augmented generation for large language models: A survey . ArXiv, abs/2312.10997

    work page internal anchor Pith review Pith/arXiv arXiv 2023

  47. [69]

    Edward J Hu, yelong shen, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, Lu Wang, and Weizhu Chen. 2022. https://openreview.net/forum?id=nZeVKeeFYf9 Lo RA : Low-rank adaptation of large language models . In International Conference on Learning Representations

    work page 2022

  48. [70]

    Yizheng Huang and Jimmy X. Huang. 2024. https://api.semanticscholar.org/CorpusID:269188036 A survey on retrieval-augmented text generation for large language models . ArXiv, abs/2404.10981

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  49. [71]

    Bernd Huber, Ghazal Fazelnia, Andreas Damianou, Sebastian Peleato, Max Lefarov, Praveen Ravichandran, Marco De Nadai, Mounia Lalmas-Roellke, and Paul N. Bennett. 2025. https://arxiv.org/abs/2505.17051 Embedding-to-prefix: Parameter-efficient personalization for pre-trained large language models . Preprint, arXiv:2505.17051

    work page arXiv 2025

  50. [72]

    Gautier Izacard, Mathilde Caron, Lucas Hosseini, Sebastian Riedel, Piotr Bojanowski, Armand Joulin, and Edouard Grave. 2022. https://api.semanticscholar.org/CorpusID:249097975 Unsupervised dense information retrieval with contrastive learning . Transactions on Machine Learning Research, 2022

    work page 2022

  51. [73]

    Omar Khattab and Matei Zaharia. 2020. https://doi.org/10.1145/3397271.3401075 Colbert: Efficient and effective passage search via contextualized late interaction over bert . In Proceedings of the 43rd International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR '20, page 39–48, New York, NY, USA. Association for Computing...

    work page doi:10.1145/3397271.3401075 2020

  52. [74]

    Patrick Lewis, Ethan Perez, Aleksandra Piktus, Vladimir Karpukhin, Naman Goyal, and 1 others. 2020. https://proceedings.neurips.cc/paper/2020/hash/6b493230205f780e1bc26945df7481e5-Abstract.html Retrieval-augmented generation for knowledge-intensive nlp . In Advances in Neural Information Processing Systems 33 (NeurIPS 2020)

    work page 2020

  53. [75]

    Xiang Lisa Li and Percy Liang. 2021. https://doi.org/10.18653/v1/2021.acl-long.353 Prefix-tuning: Optimizing continuous prompts for generation . In Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers), pages 4582--4597, Onl...

    work page doi:10.18653/v1/2021.acl-long.353 2021

  54. [76]

    Xiaopeng Li, Pengyue Jia, Derong Xu, Yi Wen, Yingyi Zhang, Wenlin Zhang, Wanyu Wang, Yichao Wang, Zhaocheng Du, Xiangyang Li, Yong Liu, Huifeng Guo, Ruiming Tang, and Xiangyu Zhao. 2025 a . https://arxiv.org/abs/2504.10147 A survey of personalization: From rag to agent . Preprint, arXiv:2504.10147

    work page arXiv 2025

  55. [77]

    Zongxi Li, Zijian Wang, Weiming Wang, Kevin Hung, Haoran Xie, and Fu Lee Wang. 2025 b . https://doi.org/10.1016/j.caeai.2025.100417 Retrieval-augmented generation for educational application: A systematic survey . Computers and Education: Artificial Intelligence, 8:100417

    work page doi:10.1016/j.caeai.2025.100417 2025

  56. [78]

    Chin-Yew Lin. 2004. https://aclanthology.org/W04-1013/ ROUGE : A package for automatic evaluation of summaries . In Text Summarization Branches Out, pages 74--81, Barcelona, Spain. Association for Computational Linguistics

    work page 2004

  57. [79]

    Ilya Loshchilov and Frank Hutter. 2019. https://arxiv.org/abs/1711.05101 Decoupled weight decay regularization . Preprint, arXiv:1711.05101

    work page internal anchor Pith review Pith/arXiv arXiv 2019

  58. [80]

    Leland McInnes, John Healy, and Sean Astels. 2017. https://doi.org/10.21105/joss.00205 hdbscan: Hierarchical density based clustering . Journal of Open Source Software (JOSS), 2(11)

    work page doi:10.21105/joss.00205 2017

  59. [81]

    Shi Na, Liu Xumin, and Guan Yong. 2010. https://doi.org/10.1109/IITSI.2010.74 Research on k-means clustering algorithm: An improved k-means clustering algorithm . In 2010 Third International Symposium on Intelligent Information Technology and Security Informatics, pages 63--67, Jian, China

    work page doi:10.1109/iitsi.2010.74 2010

  60. [82]

    Lin Ning, Luyang Liu, Jiaxing Wu, Neo Wu, Devora Berlowitz, Sushant Prakash, Bradley Green, Shawn O'Banion, and Jun Xie. 2025. https://doi.org/10.1145/3701716.3715463 User-llm: Efficient llm contextualization with user embeddings . In Companion Proceedings of the ACM on Web Conference 2025, WWW '25, page 1219–1223, New York, NY, USA. Association for Compu...

    work page doi:10.1145/3701716.3715463 2025

  61. [83]

    Adam Paszke, Sam Gross, Francisco Massa, Adam Lerer, James Bradbury, Gregory Chanan, Trevor Killeen, Zeming Lin, Natalia Gimelshein, Luca Antiga, and 1 others. 2019. https://dl.acm.org/doi/10.5555/3454287.3455008 Pytorch: An imperative style, high-performance deep learning library . Advances in neural information processing systems, 32

    work page doi:10.5555/3454287.3455008 2019

  62. [84]

    Yilun Qiu, Tianhao Shi, Xiaoyan Zhao, Fengbin Zhu, Yang Zhang, and Fuli Feng. 2025. https://doi.org/10.18653/v1/2025.emnlp-main.536 Latent inter-user difference modeling for LLM personalization . In Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing, pages 10610--10628, Suzhou, China. Association for Computational Linguistics

    work page doi:10.18653/v1/2025.emnlp-main.536 2025

  63. [85]

    Stephen Robertson, Steve Walker, Susan Jones, Micheline Hancock-Beaulieu, and Mike Gatford. 1995. https://www.researchgate.net/publication/221037764_Okapi_at_TREC-3 Okapi at trec-3. In In Proceedings of the Third Text REtrieval Conference, pages 109--126, Gaithersburg, MD: NIST. TREC-3

    work page arXiv 1995

  64. [86]

    APACrefauthors \ 1987

    Peter J. Rousseeuw. 1987. https://doi.org/10.1016/0377-0427(87)90125-7 Silhouettes: A graphical aid to the interpretation and validation of cluster analysis . Journal of Computational and Applied Mathematics, 20:53--65

    work page doi:10.1016/0377-0427(87)90125-7 1987

  65. [87]

    Pranab Sahoo, Ayush Kumar Singh, Sriparna Saha, Vinija Jain, Samrat Mondal, and Aman Chadha. 2025. https://arxiv.org/abs/2402.07927 A systematic survey of prompt engineering in large language models: Techniques and applications . Preprint, arXiv:2402.07927

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  66. [88]

    Alireza Salemi, Surya Kallumadi, and Hamed Zamani. 2024 a . https://doi.org/10.1145/3626772.3657783 Optimization methods for personalizing large language models through retrieval augmentation . In Proceedings of the 47th International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR '24, page 752–762, New York, NY, USA. Ass...

    work page doi:10.1145/3626772.3657783 2024

  67. [89]

    Alireza Salemi, Sheshera Mysore, Michael Bendersky, and Hamed Zamani. 2024 b . https://doi.org/10.18653/v1/2024.acl-long.399 L a MP : When large language models meet personalization . In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 7370--7392, Bangkok, Thailand. Association for Comp...

    work page doi:10.18653/v1/2024.acl-long.399 2024

  68. [90]

    Keshav Santhanam, Omar Khattab, Jon Saad-Falcon, Christopher Potts, and Matei Zaharia. 2022. https://doi.org/10.18653/v1/2022.naacl-main.272 C ol BERT v2: Effective and efficient retrieval via lightweight late interaction . In Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language ...

    work page doi:10.18653/v1/2022.naacl-main.272 2022

  69. [91]

    Kiriakos Sgardelis, Dionisis Margaris, Dimitris Spiliotopoulos, and Costas Vassilakis. 2025. https://doi.org/10.1007/s41060-025-00846-4 An evaluation review of user similarity metrics in sparse collaborative filtering datasets . International Journal of Data Science and Analytics, 20:6665--6693

    work page doi:10.1007/s41060-025-00846-4 2025

  70. [92]

    Chenglei Shen, Xiao Zhang, Teng Shi, Changshuo Zhang, Guofu Xie, and Jun Xu. 2024. https://api.semanticscholar.org/CorpusID:271050125 A survey of controllable learning: Methods and applications in information retrieval . ArXiv, abs/2407.06083

    work page arXiv 2024

  71. [93]

    Teng Shi, Zihua Si, Jun Xu, Xiao Zhang, Xiaoxue Zang, Kai Zheng, Dewei Leng, Yanan Niu, and Yang Song. 2024. https://doi.org/10.1145/3626772.3657811 Unisar: Modeling user transition behaviors between search and recommendation . In Proceedings of the 47th International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR '24, pa...

    work page doi:10.1145/3626772.3657811 2024

  72. [94]

    Teng Shi, Jun Xu, Xiao Zhang, Xiaoxue Zang, Kai Zheng, Yang Song, and Han Li. 2025. https://doi.org/10.1145/3726302.3730075 Retrieval augmented generation with collaborative filtering for personalized text generation . In Proceedings of the 48th International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR '25, page 1294–1...

    work page doi:10.1145/3726302.3730075 2025

  73. [95]

    Shamane Siriwardhana, Rivindu Weerasekera, Elliott Wen, Tharindu Kaluarachchi, Rajib Rana, and Suranga Nanayakkara. 2023. https://doi.org/10.1162/tacl_a_00530 Improving the domain adaptation of retrieval augmented generation ( RAG ) models for open domain question answering . Transactions of the Association for Computational Linguistics, 11:1--17

    work page doi:10.1162/tacl_a_00530 2023

  74. [96]

    Peter Steinberger. 2025. https://github.com/openclaw/openclaw Openclaw: Open-source autonomous ai agent . Originally released as Clawdbot; later renamed to OpenClaw

    work page 2025

  75. [97]

    Jiakai Tang, Sunhao Dai, Teng Shi, Jun Xu, Xu Chen, Wen Chen, Wu Jian, and Yuning Jiang. 2025. https://api.semanticscholar.org/CorpusID:277435679 Think before recommend: Unleashing the latent reasoning power for sequential recommendation . ArXiv, abs/2503.22675

    work page arXiv 2025

  76. [98]

    Xiang Wang, Xiangnan He, Meng Wang, Fuli Feng, and Tat-Seng Chua. 2019. https://doi.org/10.1145/3331184.3331267 Neural graph collaborative filtering . In Proceedings of the 42nd International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR'19, page 165–174, New York, NY, USA. Association for Computing Machinery

    work page doi:10.1145/3331184.3331267 2019

  77. [99]

    Thomas Wolf, Lysandre Debut, Victor Sanh, Julien Chaumond, Clement Delangue, Anthony Moi, Pierric Cistac, Tim Rault, Remi Louf, Morgan Funtowicz, Joe Davison, Sam Shleifer, Patrick von Platen, Clara Ma, Yacine Jernite, Julien Plu, Canwen Xu, Teven Le Scao, Sylvain Gugger, and 3 others. 2020. https://doi.org/10.18653/v1/2020.emnlp-demos.6 Transformers: Sta...

    work page doi:10.18653/v1/2020.emnlp-demos.6 2020

  78. [100]

    Shitao Xiao, Zheng Liu, Peitian Zhang, Niklas Muennighoff, Defu Lian, and Jian-Yun Nie. 2024. https://doi.org/10.1145/3626772.3657878 C-pack: Packed resources for general chinese embeddings . In Proceedings of the 47th International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR '24, page 641–649, New York, NY, USA. Assoc...

    work page doi:10.1145/3626772.3657878 2024

  79. [101]

    Haoran Xin, Ying Sun, Chao Wang, and Hui Xiong. 2025. https://doi.org/10.1145/3715099 Llmcdsr: Enhancing cross-domain sequential recommendation with large language models . ACM Transactions on Information Systems, 43(5)

    work page doi:10.1145/3715099 2025

  80. [102]

    Fangyuan Xu, Weijia Shi, and Eunsol Choi. 2023. https://api.semanticscholar.org/CorpusID:263830734 Recomp: Improving retrieval-augmented lms with compression and selective augmentation . ArXiv, abs/2310.04408

    work page arXiv 2023

Showing first 80 references.